Moxie Marlinspike, co-founder of encrypted messaging platform Signal, has launched Confer, a new artificial intelligence startup positioning itself as a privacy-first alternative to mainstream AI assistants such as ChatGPT and Claude.
Marlinspike says Confer offers a chat-based AI service that looks and functions like existing conversational tools but is designed to prevent the collection, storage, or reuse of user data.
Unlike most AI assistants, Confer’s conversations cannot be used to train models or target advertising because the company’s infrastructure is designed so that it never has access to user chats.
“It’s a form of technology that actively invites confession,” says Marlinspike. “Chat interfaces like ChatGPT know more about people than any other technology before. When you combine that with advertising, it’s like someone paying your therapist to convince you to buy something.”
How Confer Keeps Conversations Private
Messages sent to and from Confer are encrypted using the WebAuthn passkey system, a standard designed to reduce reliance on traditional passwords. While WebAuthn works most seamlessly on mobile devices and Apple Macs running macOS Sequoia, it can also be used on Windows or Linux systems with the help of password managers.
On the server side, Confer processes all AI inference inside a Trusted Execution Environment (TEE). This hardware-based security model is designed to isolate sensitive computations from the rest of the system, even from system administrators. Remote attestation mechanisms are also in place, allowing verification that the environment has not been tampered with.
Within that secure enclave, an array of open-weight foundation models handles user queries. Because the system is architected so that decrypted data never leaves the protected environment, Confer’s operators cannot log, inspect or store conversations, even if they wanted to.
The result is a far more complex setup than standard AI inference pipelines, which are already technically demanding. However, proponents argue that the added complexity is necessary to ensure users can discuss sensitive topics without fear of surveillance or data leakage.
A Different Business Model for AI
Confer’s privacy-first approach is reflected in its pricing. The free tier is limited to 20 messages per day and five active chats. Users who pay $35 a month gain unlimited access, alongside more advanced models and personalisation features.
That price point is notably higher than OpenAI’s ChatGPT Plus subscription, underscoring a central tension in the AI market: services that do not rely on data harvesting or advertising must charge more to remain sustainable.
“Privacy doesn’t come cheap,” Marlinspike has argued, suggesting that users may need to make a conscious trade-off between cost and control over their personal information.
Positioning in a Crowded AI Market
Confer enters a competitive AI assistant market dominated by OpenAI, Anthropic and Google. However, the startup is differentiating itself by focusing on users concerned about data retention, AI training practices and advertising-driven business models.
By building privacy protections directly into its technical architecture rather than relying on policy promises, Confer is betting that a segment of users will pay a premium for AI services that do not monetise personal data.
Whether privacy-first AI assistants can compete with better-funded, data-rich rivals remains an open question. But as AI systems become more intimate and deeply embedded in everyday decision-making, experts say Confer’s launch highlights a growing demand for tools that treat user privacy not as a feature, but as a foundation.
Talking Points
It is notable that Confer is attempting to solve one of the biggest trust gaps in consumer AI by designing its assistant to avoid data collection altogether, rather than relying on privacy policies or user promises.
By encrypting conversations end-to-end and running AI inference inside secure hardware environments, Confer positions itself as a practical alternative for users who want the benefits of AI without surrendering sensitive personal information.
At Techparley, we see Confer as part of a growing pushback against advertising-driven AI models, particularly as large AI companies begin to explore monetisation strategies that rely on deeper user profiling.
The use of open-weight models and a privacy-preserving backend suggests Confer is prioritising architectural safeguards over scale, signalling a deliberate trade-off between rapid growth and user trust.
As Confer matures, there is an opportunity to expand its relevance through enterprise, developer, or regulated-sector use cases where data protection is not optional but mandatory. If executed well, Confer could set a reference point for how privacy-first AI services are built and priced.
——————-
Bookmark Techparley.com for the most insightful technology news from the African continent.
Follow us on Twitter @Techparleynews, on Facebook at Techparley Africa, on LinkedIn at Techparley Africa, or on Instagram at Techparleynews.
